Multi-Channel Wireless Networks: Capacity and Protocols

1
Multi-Channel Wireless Networks Capacity and
Protocols

• Pradeep Kyasanur and Nitin H. Vaidya
• University of Illinois at Urbana-Champaign

2
Wireless networks

• We consider multi-hop networks
• Ad hoc networks, mesh networks, sensor networks

3
Key limitation

• Wireless channel is a shared resource
• Simultaneous transmissions limited by
  interference
• Throughput reduces with multiple hops
• Higher density reduces per-node throughput
• Throughput reduces as number of flows increase
• New applications require higher throughput
• Streaming video, games
• Improving network capacity is important

4
Multiple channels

• Typically, available frequency spectrum is split
  into multiple channels
• Large number of channels may be available
• Using all the available channels is beneficial

3 channels
8 channels
4 channels
26 MHz
100 MHz
200 MHz
150 MHz
2.45 GHz
915 MHz
5.25 GHz
5.8 GHz
250 MHz
500 MHz
1000 MHz
61.25 GHz
24.125 GHz
122.5 GHz
5
Current state of art

• Typical multi-hop networks use one channel only
• Key challenge Connectivity vs using multiple
  channels

6
Multiple interfaces

• Nodes may be equipped with multiple interfaces
• Common case may be small number of interfaces
• Wireless radio interfaces typically support one
  channel at a time
• We assume a half-duplex transreceiver
• Interface can switch to any channel
• Number of interfaces per node expected to be
  smaller than number of channels

7
Example configuration

• IEEE 802.11 has multiple channels
• 12 in IEEE 802.11a
• Devices can be equipped with multiple interfaces
• E.g., one interface per PCMCIA/ mini-PCI slot
• Typically, fewer interfaces than channels
• 2 interfaces, 12 channels

8
Focus of research

• Establish capacity of multi-channel networks
• How does capacity vary with channels?
• What are the insights from theoretical study?
• Design, implement and evaluate protocols
• Can we use existing protocols?
• Develop suitable protocols optimized for
  multi-channel networks
• How to implement protocols in real systems?

9
Organization

• Capacity analysis
• Theory to protocols Overview of challenges
• Protocols
• Interface Management Protocol
• Routing Protocol
• Implementation Issues
• Summary and Future Work

10
Capacity problem

• Per-node capacity decreases as network density
  increases
• Use more channels when network density increases
• Challenge Harder to scale interfaces at the same
  rate as channels
• How does the network capacity scale with large
  number of channels, and fewer interfaces than
  channels?

11
Related work

• GuptaKumar have studied the capacity of single
  channel networks
• Result applicable for multi-channel networks when
  number of channels number of interfaces per
  node
• Gamal et al. have studied the throughput-delay
  tradeoff
• Some of our constructions are based on their work
• Lot of work on studying capacity in other
  contexts
• Mobility, infrastructure-support, delay
  constraints, etc.

12
Model

• n nodes in the network, all located on a unit
  torus
• c channels are available
• m interfaces per node
• Interface operates on one channel at a time
• Channel model 1 Total bandwidth W, each channel
  has bandwidth W/c
• Channel model 2 Total bandwidth Wc, each channel
  has bandwidth W

13
Network scenarios GuptaKumar

• Arbitrary network
• Nodes can be located anywhere on the torus
• Traffic patterns can be arbitrarily chosen
• Measure of capacity aggregate network transport
  capacity (bit-meters/sec)
• Random network
• Nodes are randomly placed on the torus
• Each node sets up a flow to a random destination
• Measure of capacity minimum of flow throughputs
  (bits/sec)

14
Results

• Established tight bounds
• Upper bounds and constructive lower bounds have
  same order
• Capacity depends on ratio of c to m
• Derived insights from constructions
• Capacity-optimal routing and scheduling strategies

15
Arbitrary network Region 1
16
Arbitrary network Region 2
17
Random network Region 1
18
Random network Region 2
19
Random network Region 3
20
Practical implications

• When m lt c, it is better to use c channels
• If only m channels are used, larger capacity loss
• Single interface per node often suffices
• Up to log(n) channels, 1 interface is sufficient
• Switching delay may not affect capacity
• Extra hardware has to be provided

21
Insights for protocol development

• Multiple interfaces can simplify protocol design
• Use one interface for receiving data on a fixed
  channel
• Use second interface for sending data
• Routing protocol has to distribute routes
• Important for multi-channel networks
• Optimal transmission range depends on density of
  nodes as well as number of channels
• Optimum of interfering nodes of channels

22
Open issues

• Impact of switching delay has to be better
  studied
• Is switching required at all?
• Capacity under other switching constraints
  switch among only a subset of channels
• Analyze capacity of deterministic networks
• Given a topology, what is the capacity?
• What protocols should be used to achieve this
  capacity?

23
Organization

• Capacity analysis
• Theory to protocols Overview of challenges
• Protocols
• Interface Management Protocol
• Routing Protocol
• Implementation Issues
• Summary and Future Work

24
Assumptions

• Homogeneous channels Channels with similar
  ranges and rates
• Possibly channels in same frequency band
• Alternatively, use appropriate power control

25
Design choice Multiple interfaces

• Theory indicates single interface may suffice
• But, multiple interfaces can hide switching delay
• Multiple interfaces simplify protocols
• Our proposal, described later, is simple to
  implement
• Multiple interfaces can allow full-duplex
  transfer
• Useful when multiple channels are available

26
Design choice Protocol separation

• Separate protocol design into two components
• Interface management
• Routing
• Interface management shorter timescales
• Map interfaces to channels
• Schedule and control interface switching
• Routing longer timescales
• Select channel diverse routes

27
Protocol separation overview
28
Link layer requirements

• Utilize all the available channels
• Even if number of interfaces lt number of channels
• E.g. Interfaces can be switched to different
  channels

• Ensure connectivity is not affected
• B should be able to communicate with A and D
• Need to be cognizant of switching delay

29
Link layer requirements

• Solution should be simple to implement
• Avoid the need for complicated co-ordination,
  tight time synchronization
• Allow implementation with existing hardware
• Avoid requiring hardware changes
• Avoid assuming specific hardware capabilities

30
Routing requirements

• Improve single flow throughput by using multiple
  channels
• Both interfaces can be utilized at the relay nodes

• Improve network throughput by distributing flows

31
Organization

• Capacity analysis
• Theory to protocols Overview of challenges
• Protocols
• Interface Management Protocol
• Routing Protocol
• Implementation Issues
• Summary and Future Work

32
Key components

• Interface assignment strategy
• How to map interfaces to channels?
• How to ensure neighboring nodes can communicate
  with each other?
• Interface management protocol
• Control when interfaces are switched, based on
  assignment strategy
• Buffer packets if interface is busy

33
Interface assignment strategies

• Static Interface Assignment
• Interface to channel assignment is fixed
• Dynamic Interface Assignment
• Interface assignment changes with time
• Hybrid Interface Assignment
• Some interfaces use static assignment, others use
  dynamic assignment

34
Static interface assignment

• Each interface is fixed to one channel
• Does not require frequent co-ordination

35
Dynamic interface assignment

• Interfaces can switch channels as needed
• E.g., So2004Mobihoc, Bahl2004Mobicom

36
Hybrid strategies

• One common channel used as control channel
• One interface always fixed to this channel
• Remaining channels used as data channels
• Second interface switches among data channels

Common control channel becomes a bottleneck
37
Proposed hybrid assignment

• One interface fixed on a channel
• Different nodes use different fixed channels
• Other interfaces switch as needed
• Dynamic assignment
• Fixed interface receives data on well-known
  channel
• Avoids co-ordination issues, deafness problems
• Switchable interfaces send on recipient's fixed
  channel
• Retain flexibility of dynamic assignment

38
Hybrid assignment example
Any node pairs within transmission range can
communicate
39
Identifying fixed channel

• Static Approach Fixed channel as a function of
  node-identifier
• Simple to build, but may not balance assignment
• Dynamic approach Choose fixed channel based on
  neighborhood information
• A node chooses least used channel for fixed
  channel
• Can balance load, and still inexpensive

40
Interface management

• Each channel is associated with a queue
• Broadcast packets are inserted in to every queue
• Fixed interface services fixed channel queue
• Switchable interface services other channels
• Channels serviced in round-robin fashion
• Each channel is serviced for at most MaxSwitchTime

41
UDP throughput chain topology
42
FTP throughput chain topology
43
Open issues

• Broadcast cost increases linearly with channels
• Consider partial broadcasts
• Use a separate broadcast channel, with third
  interface
• Fixed channel selection is topology-based
• Consider load, channel quality information
• Integrate with a routing solution

44
Organization

• Capacity analysis
• Theory to protocols Overview of challenges
• Protocols
• Interface Management Protocol
• Routing Protocol
• Heterogeneous channels
• Summary and Future Work

45
Routing approach

• Existing routing protocols can be operated over
  interface management protocol
• May not select channel diverse routes
• Does not consider cost of switching interfaces
• Our solution
• Develop a new channel-aware metric
• Incorporate metric in an on-demand source-routed
  protocol

46
Selecting channel diverse routes

• Most routing protocols use shortest-hop metric
• Not sufficient with multi-channel networks
• Need to exploit channel diversity

47
Impact of switching cost

• Interface switching cost has to be considered
• Switching interfaces incurs a delay
• A node may be on different routes, requiring
  switching

48
Designing a routing metric

• Measure switching cost for a channel
• Measure total link cost of a hop
• Combine individual link costs into path cost

49
Measuring switching cost

• Switching cost depends on the likelihood a switch
  is necessary before transmission
• Fixed channel has cost 0
• Active channel has low switching cost
• Switching cost (SC) directly proportional to time
  spent on other channels

50
Routing protocol

• Incorporate metric in on-demand source-routed
  protocol (similar to DSR)
• RREQ messages modified to include link costs
• Source initiates RREQ
• Intermediate nodes forward RREQ if,
• New RREQ
• Cost of RREQ smaller than previously seen RREQ
• Destination can compute best path
• Using link cost information in sent RREQ

51
Throughput in random networks
52
Throughput with varying load
53
Open issues

• Incorporate load information into MCR metric
• Support for route caching
• Metric does not allow route combination
• Design alternate metrics?
• Integrated routing and fixed channel selection
• Can improve performance at cost of increased
  complexity

54
Organization

• Capacity analysis
• Theory to protocols Overview of challenges
• Protocols
• Interface Management Protocol
• Routing Protocol
• Implementation Issues
• Summary and Future Work

55
Lack of multi-channel support

• Existing assumptions break with multiple channels
• Assume of channels of interfaces
• Routing table has interface information only
• Not easy to use multiple interfaces
• Switching channels requires explicit invocation
• Interfaces and channels not hidden from
  applications
• Frequent switching not permitted

56
Requirements

• Hide interface management from data path
• Allow existing applications to work unmodified
• Break node-channel mapping
• Allow channel to be selected based on destination
• Support multi-channel / single channel broadcast
• Broadcast primitive required for many applications

57
Proposed architecture

• Abstraction layer exports single virtual
  interface
• Channel switching details are hidden
• Fixed channel selection, and routing protocol is
  implemented as part of channel policy manager

Joint work with Chandrakanth Chereddi
58
Organization

• Capacity analysis
• Theory to protocols Overview of challenges
• Protocols
• Interface Management Protocol
• Routing Protocol
• Heterogeneous channels
• Implementation Issues
• Summary and Future Work

59
Summary

• Goal of the project is to utilize multiple
  channels
• Research issues considered are
• Analysis of capacity of multi-channel networks
• Design of protocols for multi-channel networks
• Implementing protocol suite in testbed

60
Future work

• Capacity analysis with switching delay
• What if there is no switching allowed at all?
• Flow-aware protocol design
• Assign channels based on channel quality and load
• Select routes based on existing routes
• Implementation and measurement
• Fully implement all protocols
• Measure characteristics of multiple channels

61
Questions?

• More details at
• http//www.crhc.uiuc.edu/wireless

62
Backup Slides
63
Arbitrary Network Upper bound

• Interference constraints GuptaKumar Each pair
  of simultaneous receivers must have minimum
  separation
• Separation depends on transmission radius
• Bounds the number of simultaneous transmissions
• Interface constraint Only m interfaces available
• Each node can send/receive at most m bits/sec

64
Arbitrary Network Lower bound

• Divide torus in to square cells
• Each cell has nodes

65
Random Networks Upper bound

• Arbitrary network constraints Random network is
  a special case of an arbitrary network
• Connectivity constraint A minimum transmission
  range is needed to ensure network is connected
• Destination bottleneck constraint The maximum
  number of incoming flows at any node will limit
  per-flow throughput

66
Lower bound Routing

• Divide torus in to square cells of area a(n)
• a(n) depends on the number of channels
• Route through cells on the straight line joining
  source and destination

67
Lower bound Step 1 schedule

• Divide every second in to hop-color slots
• Flow scheduling For each hop of a flow, schedule
  its transmission in some hop-color slot
• Procedure
• Build a routing graph
• Vertices are nodes in the network
• One edge for every hop
• Edge color the graph
• Number of colors used number of hop-color slots
• Map each color to a hop-color slot
• Every hop is scheduled in slot associated with
  its color

68
Lower bound Step 2 schedule

• Divide each hop-color slot in to node slots
• Node scheduling Each node can only transmit in
  its node slot
• Procedure
• Build an interference graph
• Vertices are nodes in the network
• One edge for every pair of nodes that may
  interfere
• Vertex color the graph
• of colors of node slots per hop-color slot
• Map each color to a slot
• Each node transmits only in slot associated with
  its color

69
Switching Delay

• Initial analysis ignores interface switching
  delay
• Upper bounds do not mandate switching
• Open question Is interface switching required at
  all
• Possible that switching delay does not affect
  capacity
• Lower bound constructions affected by delay
• Capacity affected only if there are latency
  constraints
• Even with latency constraints, multiple
  interfaces can hide delay

70
Benefits of Proposed Strategy

• Frequent co-ordination not required
• Fixed channel information infrequently exchanged
• Maintains full-connectivity
• Any node pairs within transmission range can
  communicate
• No changes required to MAC protocol
• Can be built with existing IEEE 802.11 hardware

71
Arbitrary networks

• Two capacity regions

72
Random networks

• Three capacity regions

73
One approach

• Based on ETT measurement Draves2004Mobicom
• ETT(j) Expected Transmission Time of packet
• LinkLossRate measurement modified
• LinkRate measured from probing driver

74
One path metric (MCR)

• Based on WCETT Draves2004Mobicom
• Path cost limited by bottleneck channel cost ( Xj
  )
• Network throughput depends on aggregate cost

75
CBR throughput
76
CBR throughput